---
dataset_info:
  features:
  - name: id
    dtype: string
  - name: seed_prompt
    dtype: string
  - name: language
    dtype: string
  - name: model
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: agent_trace
    list: json
  - name: research_early_stopping
    dtype: bool
  - name: source_id
    dtype: string
  splits:
  - name: train
    num_bytes: 110358463
    num_examples: 4986
  download_size: 108068951
  dataset_size: 110358463
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---